Abstract
Purpose:
To report reference centile risk curves of cross-sectional data from the Guangzhou RESC study, and then evaluate the predictive power of the references using longitudinal data from the Guangzhou Twin Eye Study.
Methods:
The cycloplegic refraction of 4218 children aged 5-15 years collected in the Refractive Error Study in Children (RESC) in urban Chinese children and 1963 children with annually follow-up data in the Guangzhou Twin Eye Study were analyzed. Age- and sex-specific percentile curves for refraction were constructed by quantile regression model using the RESC data. Linear mixed model was used to access the progression of refraction in longitudinal data of twins. The risk to develop high myopia was estimated by Cox proportional hazards model. To further explore the underlying predictive factors, the myopic relevant characteristics of the two groups were also described.
Results:
Age and sex adjusted smoothed reference centile curves for the 97th, 90th, 75th, 50th, 25th, 10th and 3rd centiles were created based on the population-based data on 2175 boys and 2043 girls. Linear mixed model showed that progression of SE was at least -3 diopters greater in those with baseline refraction under 3rd, 5th and 10th centiles( p<0.01) compared to the other children. In Cox proportional hazard model, odds ratios for children below 3rd, 5th and 10th centiles to develop high myopia were 50.61, 47.82 and 45.51 respectively (for each p<0.01). In the comparison of myopia-relevant characteristics, parental myopia showed significant difference between the two groups using the reference centiles (for each p<0.05).
Conclusions:
The reference centile curves provide information on the age- and sex-specific distribution of refraction, and can be used to identify the children on the more myopic extreme at their age group who may be at increased risk of developing high myopia by adulthood.
Keywords: 605 myopia •
676 refraction •
497 development